WO2016075827A1

WO2016075827A1 - Recommendation system, recommendation method and recommendation program

Info

Publication number: WO2016075827A1
Application number: PCT/JP2014/080259
Authority: WO
Inventors: ロビンスウィジー
Original assignee: 楽天株式会社
Priority date: 2014-11-14
Filing date: 2014-11-14
Publication date: 2016-05-19
Also published as: JP6229074B2; JPWO2016075827A1

Abstract

This recommendation system recommends plans with an associated place and date to a user, and is provided with an acquisition means which acquires a user ID identifying a user, an extraction means which, on the basis of the history of selection of plans by the user identified by the user ID, determines at least one of multiple methods for extracting plans to recommend to the user and extracts a plan using the determined method, and a display control means which displays the plan information extracted by the extraction means.

Description

Recommendation system, recommendation method and recommendation program

One aspect of the present invention relates to a recommendation system, a recommendation method, and a recommendation program.

An information system that reserves and settles a plan for golf, travel, etc. via the Internet is known (for example, see Patent Document 1). In addition, in a reservation service such as a golf course, there is a plan that presents a plan according to a campaign on the business side or presents a plan determined as an advertisement.

JP 2010-15493 A

Since the presentation of a plan as an advertisement is the same plan for all users, it may not be suitable for the users. Therefore, it is desirable to present a more suitable plan to the user.

On the other hand, methods such as collaborative filtering and regional filtering can be used to extract plans to recommend to users, but it is not known which method will be used to extract suitable plans for users. There are cases where a plan extracted using is not suitable for the user.

Accordingly, an object of one aspect of the present invention is to provide a system capable of presenting a suitable plan that meets a user's wishes to the user.

In order to solve the above problems, a recommendation system according to an aspect of the present invention is a recommendation system that recommends a plan to a user, and is specified by an acquisition unit that acquires a user ID that specifies the user, and the user ID. Extracting means for extracting at least one of a plurality of methods for extracting a plan to be recommended to the user based on a history of plan selection by the user, and extracting the plan using the determined method; Display control means for displaying the plan information of the selected plan.

A recommendation method according to an aspect of the present invention is a recommendation method in a recommendation system for recommending a plan to a user, an acquisition step of acquiring a user ID for specifying the user, and a plan selection by the user specified by the user ID Based on the history, the at least one method among a plurality of methods for extracting a plan to recommend to the user is determined, and a plan is extracted using the determined method, and the plan information of the extracted plan And a display control step for displaying.

A recommendation program according to an aspect of the present invention is a recommendation program for causing a computer to function as a recommendation system for recommending a plan to a user. The acquisition program acquires a user ID for identifying the user in the computer, and a user. Extraction that determines at least one of a plurality of methods for extracting a plan to be recommended to the user based on the plan selection history by the user specified by the ID, and extracts the plan using the determined method And a display control function for displaying the plan information of the extracted plan.

According to the above aspect, a suitable method for extracting a plan for recommending to a user is determined based on a user's plan selection history, and the plan extracted using the determined method is determined to the user. To be presented. Therefore, the user can obtain plan information of a plan that meets his / her wishes.

In the recommendation system according to another aspect, the extraction unit includes a calculation unit that calculates a score for each plan based on the user's history, and extracts the plan based on the score calculated by the calculation unit.

According to the above aspect, the score for each plan is calculated based on the history related to the user's plan selection, and the plan information is presented to the user based on the calculated score. The score calculated based on the user's history is highly likely to reflect the degree of user's desire. Since the plan information is presented to the user based on such a score, the user can obtain the plan information of the plan that meets his / her wish.

In the recommendation system according to another aspect, the calculating unit acquires, in the history, one plan information selected in the past as reference plan information, and reference location information for specifying a location associated with the plan of the reference plan information Based on the information, the plan information storage means storing the plan information is referred to, and a place information candidate that is a candidate to be recommended to the user is obtained by a predetermined method, and is associated with the obtained place information candidate. Plan information is acquired as a plan information candidate recommended to the user, and for each acquired plan information candidate, a score is calculated based on at least one of history and reference plan information and the plan information candidate attribute. Also good.

According to the above aspect, since the location information candidates are acquired based on the location information specified by the reference plan information acquired from the user history, it is possible to obtain the location information candidates that have a high probability of meeting the user's wishes. Then, for each plan information candidate associated with the location information candidate, a score is calculated based on the attribute of the user's history or reference plan information and the attribute of the plan information candidate. A score with a high probability of being reflected is obtained. By presenting plan information to the user based on such a score, the user can obtain plan information that meets his / her wishes.

In the recommendation system according to still another aspect, the calculating means may acquire the plan information of the latest selected plan in the history as reference plan information.

According to this aspect, plan information having a high probability of having attributes that meet the user's wishes is acquired as reference plan information. The candidate location information acquired based on such reference plan information and the calculated score appropriately reflect the user's wishes.

In the recommendation system according to another aspect, the calculation unit acquires location information of a location satisfying a predetermined condition among locations associated with the plan included in the history as a location information candidate, and is acquired based on the history. Location information candidates are acquired based on at least one of collaborative filtering based on the user's history and geographical information associated with the user when the number of candidate location information is less than a predetermined number. It is good to do.

According to this aspect, the location information candidates are acquired using the user history preferentially, and therefore the user history is more strongly reflected in the acquired location information candidates. Therefore, a candidate for location information having a high probability of meeting the user's wishes can be obtained.

In the recommendation system according to another aspect, when the number of selected plans is greater than or equal to a predetermined number in the user's history, the calculating means is a predetermined one of the locations associated with the plan included in the history. If the number of selected plans is less than a predetermined number in the user history, the location information of the location satisfying the above condition is acquired as a location information candidate, and is associated with the user based on collaborative filtering based on the user history. The candidate location information may be acquired based on at least one of the geographical information.

According to this aspect, when the number of plans in the user's history is small, the history may not properly reflect the user's wishes, but when the number of plans in the history is a predetermined number or more, Since the location information candidate is acquired based on the location information associated with the plan included in the user history, the user's desire is appropriately reflected in the acquired location information candidate. On the other hand, when the number of plans in the user's history is less than a predetermined number, location information candidates are acquired based on collaborative filtering or the user's geographical information. can get.

In the recommendation system according to another aspect, the calculation unit acquires the location information candidates by each of a plurality of predetermined methods, and acquires the plan information candidates associated with the location information candidates for each predetermined method, And the score of each plan information candidate is calculated, and the extraction means performs the extraction of the plan from the plan information candidates based on the calculated score for each predetermined method, and the extracted plan The predetermined method having the highest degree of matching with the plan included in the history may be determined as a method for extracting a plan recommended to the user.

最適 Based on the degree of coincidence between the plan extracted using each method and the plan included in the history, an optimal method for acquiring location information candidates is selected. As a result, a method having a high probability of extracting a plan that meets the user's wish is determined as a method for extracting a plan recommended to the user. Thereby, the probability that the plan which suits a user's hope is extracted is improved.

In the recommendation system according to another aspect, the calculating unit calculates a score of the plan information candidate based on the similarity between the attribute of the plan information candidate and the attribute of the history or reference plan information, and the calculated score is It may be higher as the degree of similarity is higher.

According to this aspect, the score is calculated based on the similarity between the user history or the attribute of the reference plan information and the attribute of the plan information candidate. The degree of matching is reflected. Since the plan information is presented to the user based on such a score, the plan information that meets the user's wish can be presented.

In the recommendation system according to another aspect, the calculating means determines whether or not there is a predetermined tendency for the location associated with the plan included in the user's history, and if it is determined that there is a predetermined tendency When a geographical parameter is used in calculating the score for each plan, it is possible to increase the weighting of the parameter, when it is not determined that there is a predetermined tendency.

According to this aspect, the weight for the geographic parameter is adjusted when calculating the score according to the presence or absence of a predetermined tendency for the location associated with the plan in the user's history. Thereby, a user's hope is more appropriately reflected with respect to the calculated score.

In the recommendation system according to another aspect, the plan information is plan information relating to golf play, and includes information on a golf course as an attribute relating to a place, and information relating to a date and time of play as an attribute relating to a date. It is good.

According to this aspect, the user can obtain plan information regarding golf play.

According to one aspect of the present invention, it is possible to provide a system that can present a plan that meets the user's wishes to the user.

It is a block diagram which shows the functional structure of the recommendation system 1 which concerns on this embodiment. It is a figure which shows the hardware constitutions of a recommendation apparatus. It is a figure which shows the example of the purchase history data of a user's golf plan. It is a figure which shows the example of a structure of the item of the plan information memorize | stored in the plan information storage part. It is a figure which shows the example and item of a user behavior data. It is a flowchart which shows the recommendation process in a recommendation system. It is a flowchart which shows the 1st example of a score calculation process. It is a flowchart which shows the example of a course ID acquisition process. It is a flowchart which shows the 2nd example of a score calculation process. It is a flowchart which shows the 3rd example of a score calculation process. It is a flowchart which shows the example of a plan information display process. It is a figure which shows the example of a display of plan information. It is a figure which shows the example of display control of plan information. It is a figure which shows the example of a display of the plan information displayed on another column. It is a figure which shows the example of a display of the plan information displayed on the calendar column. It is a figure which shows the example of a display of the plan information displayed based on the purchase history of a user. It is a figure which shows the recommendation program for functioning a computer as a recommendation system.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same or equivalent elements are denoted by the same reference numerals, and redundant description is omitted.

FIG. 1 is a block diagram showing a functional configuration of a recommendation system 1 according to the present embodiment. The recommendation system 1 of this embodiment is a system that recommends a plan to a user. The plan may be a target of browsing and purchase by the user via the Internet, and at least information related to the date may be associated, and further information related to the location may be associated. Specifically, the plan is, for example, a golf plan or a travel plan. The golf plan is associated with the date of play as information relating to the date, and the information relating to the golf course is associated as information relating to the place. In the travel plan, the travel schedule is associated with information related to the date, and the information related to the destination of the travel is correlated as information related to the place. This embodiment will be described with an example of a golf plan.

The golf plan includes, for example, course information and date / time information regarding the location of the golf course. In addition, the plan includes information such as fees and other conditions. A conventional information system that reserves and settles a golf plan via the Internet, for example, accepts an input specifying a golf course from a user and presents a plan associated with the designated golf course.

On the other hand, some information systems that sell products and the like via the Internet make recommendations by collaborative filtering. In such an information system, usage histories of products purchased or browsed by the user are stored, products based on the user's preferences and trends are extracted based on the usage history, and presented to the user To do. In a system for presenting a plan such as a conventional golf, such a recommendation by collaborative filtering can be applied. For example, a golf course that another user who has been to a golf course that the user has been to has performed can be presented to the user.

However, even if the golf course is presented to the user and the selection of the golf course from the user is accepted, there are many plans associated with the golf course, and it is difficult to find a plan that meets the user's wishes. There was a case. In addition, when there is no plan that matches the user's wish among the plans associated with the selected golf course, the user needs to perform the operation of selecting the golf course again, which is complicated.

Also, in order to recommend a plan such as golf, it is conceivable to extract a plan to recommend by collaborative filtering. However, in collaborative filtering, it is necessary to find a certain degree of correlation between a plurality of products purchased and viewed by the user, and products such as plans are sold to the user due to the characteristics of the products. Because the period is short, correlation between plans is difficult to form. Therefore, it is difficult to recommend an appropriate plan by collaborative filtering.

The recommendation system 1 according to the present embodiment solves such a problem and enables the user to present a plan that meets the user's wishes.

As shown in FIG. 1, the recommendation system 1 of the present embodiment includes a recommendation device 10 and a terminal 30. The terminal 30 is a user terminal, and can communicate with the recommendation device 10 via a network N such as the Internet. The recommendation device 10 can access the plan information storage unit 21 and the user information storage unit 22.

The recommendation device 10 functionally includes an acquisition unit 11 (acquisition unit), a calculation unit 12 (extraction unit, calculation unit), and a transmission unit 13 (display control unit). The terminal 30 includes a user ID transmission unit 31, a plan information acquisition unit 32 (display control unit), a display control unit 33 (display control unit), and a reception unit 34.

The plan information storage unit 21 is a storage unit that stores plan information. The plan information storage unit 21 may be configured as one functional unit of the recommendation device 10. The plan information storage unit 21 may be configured as a computer device that can communicate with the recommendation device 10 via a network. Details of the plan information will be described later.

The user information storage unit 22 is a storage unit that stores a user's plan purchase history as user information. The user information storage unit 22 may be configured as one function unit of the recommendation device 10. The user information storage unit 22 may be configured as a computer device that can communicate with the recommendation device 10 via a network. Details of the plan information will be described later. Details of the purchase history of the plan as user information will be described later.

FIG. 2 is a hardware configuration diagram of the recommendation device 10. As shown in FIG. 2, the recommendation device 10 is physically composed of a CPU 101 constituted by a processor, a main storage device 102 constituted by a memory such as a RAM and a ROM, an auxiliary storage device 103 constituted by a hard disk, etc., a network The computer system includes a communication control device 104 including a card, an input device 105 such as a keyboard and mouse as input devices, an output device 106 such as a display, and the like. When the recommendation device 10 is configured as a server, the recommendation device 10 may not include the input device 105 and the output device 106.

Each function of the recommendation device 10 shown in FIG. 1 is executed under the control of the CPU 101 by causing a predetermined computer software (recommendation program) to be read on hardware such as the CPU 101 and the main storage device 102 shown in FIG. This is realized by operating the communication control device 104, the input device 105, and the output device 106, and reading and writing data in the main storage device 102 and the auxiliary storage device 103. Data and databases necessary for processing are stored in the main storage device 102 and the auxiliary storage device 103.

The terminal 30 is also a computer system having the same hardware configuration as that of the recommendation device 10. The terminal 30 is configured as a device such as a personal computer or a smartphone, for example.

Subsequently, the functional unit of the recommendation device 10 will be described. The acquisition unit 11 is a part that acquires a user ID that identifies a user. Specifically, the acquisition unit 11 acquires the user ID transmitted from the terminal 30.

The calculation unit 12 is a part that calculates a score for each plan based on the plan selection history by the user specified by the user ID. The user selects a plan for purchasing or viewing the plan. Accordingly, the plan selection history is a history regarding at least one of plan purchase and browsing. Details of the score calculation will be described later. In the present embodiment, the calculation unit 12 calculates a score with reference to a golf plan purchase history. Since the plan browsing history includes plan information in the same manner as the purchase history, the score for each plan may be calculated based on the browsing history, similar to the calculation of the score based on the purchase history.

FIG. 3 is a diagram illustrating an example of purchase history data of a user's golf plan. The user information storage unit 22 stores the purchase history of the user's golf plan, and the calculation unit 12 can acquire the purchase history of the user based on the user ID by accessing the user information storage unit 22. As illustrated in FIG. 3, the user information storage unit 22 stores, for example, a plan ID, a play date, a course ID, a plan fee, and the like as a user purchase history for each user ID. The plan ID is identification information for identifying a golf plan. The play date is information on the date on which the golf plan was played. The course ID is identification information for identifying the golf course according to the golf plan. The plan fee is a fee for the golf plan.

The detailed content of the golf plan is stored in the plan information storage unit 21. FIG. 4 is a diagram illustrating an example of the configuration of the items of plan information stored in the plan information storage unit 21. The plan information includes various types of information indicating the contents of the plan in association with the plan ID. The course ID is information for identifying a golf course. The plan name is the name of the plan. The plan fee is a fee for purchasing the plan. The public period start and public period end are information indicating the start date and the end date of the period in which the plan is implemented. The day of the week classification and the target day of the week are information indicating the day of the week to which the plan is applied. The target start time zone is information indicating a time zone in which play of the plan can be started. That is, the start of the public period, the end of the public period, the day of the week classification, the target day of the week, and the target start time zone correspond to data indicating attributes regarding the date of the plan.

The plan information further includes information on other options. Lunch is information indicating whether lunch is included in the plan. The caddy flag is information indicating whether a caddy is included in the plan. The cart code is information indicating the type of cart used when playing. The number of players (minimum) and the number of players (maximum) are information indicating the lower limit and the upper limit of a set of players when playing with the plan. In addition, the plan information includes information on 4-sum discount, 2-sum guarantee, option 1 and option 2, and the like.

The calculation unit 12 can acquire the content of the golf plan purchased by the user by referring to the plan information using the plan ID included in the purchase history of the user as a key. And the calculation part 12 can produce | generate the user behavior data which shows a user's purchase history based on a user's purchase history and plan information. FIG. 5 is a diagram illustrating an example of items and values of user behavior data. The user behavior data is data obtained by converting a user purchase history into a format suitable for information processing in order to perform information processing related to the user purchase history. The user behavior data may be generated as multidimensional vector data including values of various items, for example. Specifically, the user behavior data includes values of various items included in the purchase history of the user, and includes an average value and a standard deviation, or a value obtained by normalizing them.

In the example shown in FIG. 5, MAX, COUNT, AVG, and STD indicate the maximum value, the count value, the average value, and the standard deviation of the values in parentheses, respectively. MAX (play_date) indicates the most recent play date. COUNT (c_id) indicates the number of course IDs included in the purchase history, and COUNT (DISTINCT c_id) indicates the number of course IDs included in the purchase history excluding duplicates. HOUR (start_time) indicates the start time of play. "lunch_flg" indicates a flag with lunch. caddie_flag indicates a flag with caddy. “play_fee” indicates a plan fee. “play_mem” indicates the number of players to play. play_day_weakend_flg is a flag indicating a weekend in the plan information. play_day_hol_flg is a flag indicating a holiday in the plan information. golfCourceLat indicates the latitude of the location of the golf course. golfCourceLon indicates the longitude of the location of the golf course.

The generation of user behavior data based on the purchase history of the user can be realized by a well-known technique. Further, the user behavior data may be generated by the calculation unit 12, or may be generated by another functional unit or another device. The example of the user behavior data shown in FIG. 5 represents the purchase history of the user, but it is possible to generate data that represents the contents of one plan in the same format as the user behavior data. That is, the data representing the contents of the plan can be represented, for example, as vector data including a plurality of items indicating the contents of the plan and the values of the items. Such data can also be realized by a known technique.

Such data facilitates various information processing related to the contents of the plan. That is, the user behavior data can also be said to indicate the features of the plan that the user prefers. Therefore, by comparing the data indicating the contents of the plan with the user behavior data, calculating the distance between the two vector data, etc. The similarity between the plan and the plan preferred by the user can be determined. Moreover, the similarity between users at the time of collaborative filtering can be determined by comparison with the user behavior data of other users or by calculating the distance between vector data.

Referring to FIG. 1 again, the transmission unit 13 is a part that transmits the plan information whose score is calculated by the calculation unit 12 to the terminal 30 together with the calculated score. The transmission unit 13 may transmit a predetermined number of plan information set in advance to the terminal 30.

Subsequently, the functional unit of the terminal 30 will be described. The user ID transmission unit 31 transmits the user ID to the recommendation device 10. For example, the user ID transmission unit 31 transmits the user ID input in the login process in the terminal 30 to the recommendation device 10.

The plan information acquisition unit 32 is a part that acquires the plan information transmitted from the recommendation device 10. Specifically, the plan information acquisition unit 32 acquires plan information associated with each score from the transmission unit 13 of the recommendation device 10.

The display control unit 33 is a part that displays the plan information acquired by the plan information acquisition unit 32 in the first display column displayed on the display unit of the terminal 30 for each date in time series. Details of the display control of the plan information by the display control unit 33 will be described later.

The reception unit 34 is a part that receives a selection input from the user for the plan information displayed by the display control unit 33. Processing related to acceptance of selection input for plan information will be described in detail later.

Next, with reference to FIGS. 6 to 11, the recommendation processing, score calculation processing, and plan information display control processing of this embodiment will be described in detail.

FIG. 6 is a flowchart showing a recommendation process in the recommendation system 1 of the present embodiment.

First, the user ID transmission unit 31 of the terminal 30 transmits the user ID to the recommendation device 10 (S1). Subsequently, the acquisition unit 11 of the recommendation device 10 acquires the user ID transmitted in step S1 (S2).

Next, the calculation unit 12 of the recommendation device 10 calculates a score for each plan based on the purchase history of the golf plan specified by the user ID (S3). FIG. 7 is a flowchart illustrating a first example of the score calculation process in step S3. The score calculation process will be described with reference to FIG.

The calculation unit 12 refers to the user information storage unit 22 based on the user ID acquired in step S2, and acquires the purchase history of the user (S11). The calculation unit 12 also refers to the plan information storage unit 21 in order to acquire plan information of the plan included in the purchase history.

Subsequently, the calculation unit 12 acquires a plan ID of one plan information purchased in the past as a reference plan (reference plan information) in the purchase history (S12). For example, the calculation unit 12 may acquire the latest plan information in the purchase history as a reference plan. Thereby, plan information having a high probability of having an attribute that meets the user's wish is acquired as a reference plan. Further, the calculation unit 12 may acquire plan information of a plan purchased at a time corresponding to the current season as a reference plan.

Next, the calculation unit 12 acquires a course ID (reference location information) associated with the plan ID of the plan acquired as the reference plan (S13). And the calculation part 12 acquires course ID (location information candidate) of the course to recommend based on course ID of a reference plan (S14). The calculation unit 12 determines at least one of a plurality of methods for acquiring the course ID based on the purchase history of the user, and acquires the course ID using the determined method. Since the plan is extracted based on the acquired course ID, determining the method for acquiring the course ID means determining the method for extracting the plan. Examples of the method for acquiring the course ID include a method based on the purchase history of the user, a method based on collaborative filtering, a method based on the golf course and the geographical information of the user, and the like. The course ID acquisition process for acquiring this course ID (location information) will be described below with reference to FIG.

FIG. 8 is a flowchart showing an example of course ID acquisition processing in step S14. The calculation unit 12 acquires a course ID of a course that satisfies a predetermined condition among courses associated with a plan included in the purchase history of the user as information on a recommended course (S21). The predetermined condition is, for example, that the course has been used a predetermined number of times or more in the user purchase history. Thereby, since the course ID of the recommended course is acquired by using the user history preferentially, the user history is more strongly reflected in the acquired course ID.

That is, since the course of the plan actually purchased by the user is highly likely to match the user's preference, the course ID of the course is preferentially used. For this reason, the predetermined number of times as a condition for acquiring the course ID may be one. However, even if a user purchases a plan that is associated with a course once, and the plan that is associated with that course is not purchased after that, the user is interested in the course once. However, there are cases where it is a course that does not meet the conditions if you look closely, or a course that you do not like very much if you actually play. On the other hand, a course purchased a plurality of times has a high possibility that the user likes and uses the course. Therefore, the predetermined number of times as a condition for acquiring the course ID may be two or more. In this way, only the course ID of a course purchased multiple times is acquired.

In the example of the process shown in FIG. 8, the number of course candidates for recommendation to the user is set in advance. Therefore, the calculation unit 12 determines whether or not the number of course IDs acquired in step S21 is less than a predetermined number set in advance (S22). If it is determined that the number of course IDs is less than the predetermined number, the process proceeds to step S23. On the other hand, when it is not determined that the number of course IDs is less than the predetermined number, the course ID acquisition process ends.

In step S23, the calculation unit 12 obtains a course ID based on at least one of collaborative filtering based on the user's history and geographical information associated with the user (S23).

Specifically, the calculation unit 12 refers to, for example, the purchase history of the user's golf plan (see FIG. 3) and associates the same course with the course associated with the plan that the user has purchased. Extract other users who have purchased the same plan. And the calculation part 12 acquires course ID of the course matched with the plan in the purchase history of another user. Further, the calculation unit 12 calculates the degree of similarity between the user and another user using the user behavior data as described with reference to FIG. The course ID of the course associated with the plan in the purchase history may be acquired.

Further, as the acquisition of the course ID based on the geographical information, specifically, the calculation unit 12 acquires the course ID of a course that is geographically close to the reference plan course (for example, located within a predetermined distance). . Further, the calculation unit 12 may acquire a course ID of a course that is geographically close to the user's location (for example, located within a predetermined distance).

The process of step S23 is repeated until the number of acquired course IDs exceeds a predetermined number. As described above, the calculation unit 12 determines any one of a plurality of methods as a method for acquiring the course ID, and performs the course ID acquisition process. Here, for example, when the predetermined number of times in the predetermined condition of step S21 is 2 and the predetermined number in step S22 is 40, it is used twice or more in the user's history out of 40 course IDs. All the course IDs of the course are extracted, and other course IDs extracted by collaborative filtering or geographical information are included. That is, the more courses that are used more than once, the greater the proportion of course IDs extracted based on the history out of 40 course IDs. If it does in this way, it can control to adjust the ratio of the method for extracting course ID according to a user's use situation.

Further, the calculation unit 12 may perform course ID acquisition processing according to the number of purchased plans in the user purchase history, instead of the processing of steps S21 to S23. That is, when the number of purchased plans is greater than or equal to a predetermined number in the purchase history of the user, the calculation unit 12 determines a predetermined condition (step S21) among courses associated with the plan included in the purchase history. If the number of purchased plans is less than the predetermined number in the user's purchase history, the course ID of the course satisfying the same condition as in the predetermined condition is acquired based on the user's history. A course ID is acquired based on at least one of collaborative filtering and geographical information associated with the user. As described above, the calculation unit 12 performs course ID acquisition processing using any one of a method based on the purchase history of the user, a method based on collaborative filtering, a method based on the golf course and the geographical information of the user, and the like. carry out.

According to such course ID acquisition processing, when the number of plans in the user's purchase history is small, the purchase history may not properly reflect the user's wishes. When the number is equal to or greater than the predetermined number, the course ID is acquired based on the location information associated with the plan included in the purchase history of the user, and thus the user's wish is appropriately reflected in the acquired course ID. It will be. On the other hand, when the number of plans in the user's purchase history is less than the predetermined number, the course ID is acquired based on collaborative filtering or the user's geographical information. can get.

In addition, the calculation unit 12 acquires a course ID based on the degree of matching between the plan extracted based on the course ID acquired using each of the plurality of methods and the plan included in the purchase history of the user. A technique for doing so may be determined. That is, the calculation unit 12 acquires a course ID by each of a plurality of predetermined methods, acquires a plan information candidate associated with the course ID for each predetermined method, calculates a score of each plan information candidate, And plan extraction from plan information candidates based on the calculated score, and a plan that recommends a predetermined method with the highest degree of matching of the extracted plan to the plan included in the history to the user It is good also as determining as a method for extracting. The calculation of the score for each plan and the extraction of the plan based on the score will be described later.

Specifically, the calculation unit 12 is based on, for example, a method based on the purchase history of the user, a method based on collaborative filtering (may be a plurality of types of collaborative filtering with different parameters and weights), a golf course, and geographical information of the user. A course ID is acquired by each method. Next, the calculation unit 12 acquires plan information associated with the course ID acquired by each method as a plan information candidate, calculates the score of the acquired plan information candidate, and becomes a candidate for a plan to be recommended. Extract plans based on score.

Then, the calculation unit 12 determines the degree of matching between the plan extracted for each method and the plan included in the purchase history of the user, and uses the method with the highest degree of matching as a method for acquiring the course ID. decide.

Thus, based on the degree of coincidence between the plan extracted using each method and the plan included in the history, the optimum method for acquiring the location information candidate is selected. As a result, a method having a high probability of extracting a plan that meets the user's wish is determined as a method for extracting a plan recommended to the user. Thereby, the probability that the plan which suits a user's hope is extracted is improved.

Further, in the recommendation system 1 of the present embodiment, a method for acquiring the course ID may be determined as follows.

For example, the calculation unit 12 refers to the purchase history of the user and acquires the number of courses having a purchase history of a plurality of times (for example, twice) among golf courses associated with the purchased plan. . Then, the calculation unit 12 calculates the ratio of the number of courses having a history of multiple purchases to the total number of courses that can be selected in the recommendation system 1, and when the calculated ratio is equal to or greater than a predetermined value, A process based on the purchase history is preferentially used to execute a process for acquiring the course ID.

The process preferentially using the method based on the purchase history of the user may be selecting a method based on the purchase history of the user as the process of acquiring the course ID, or in the case of using a plurality of methods, The weighting of the method based on the purchase history of the user may be made heavier than other methods. The meaning of the process that preferentially uses one method is the same as that described above in the description of the determination of the method for acquiring the following course ID.

Further, the calculation unit 12 refers to the purchase history or browsing history of the user, and uses a certain method among a plurality of methods with respect to the total number of plans for which selection input has been performed for purchase or browsing. Calculate the percentage of the number of plans that were selected and entered for purchase or viewing with respect to the recommended plan, and if the calculated percentage is greater than or equal to a predetermined value, use the method preferentially, It is good also as implementing the process which acquires course ID.

Also, for example, the calculation unit 12 acquires the page currently displayed on the user terminal 30. If the displayed page is a search screen for searching for a plan, the calculation unit 12 has a high probability that the user desires a plan or course that is not included in his / her history. The course ID may be acquired by preferentially using a method other than the method based on the purchase history of the user. On the other hand, if the displayed page is a page that displays information related to the user's history such as a so-called my page, the user wants the plan or course included in his / her history. Therefore, the calculation unit 12 may preferentially use a method based on the purchase history of the user to perform the process of acquiring the course ID.

Also, the calculation unit 12 acquires whether the current time corresponds to a so-called busy period, a non-busy period, or a quiet period based on the current date. Correspondence between each date and a busy period, a non-busy period, and a quiet period is set in advance. In view of the fact that the user tends to use a course that the user has used in the past during the busy season, the calculation unit 12 uses a method based on the purchase history of the user when the current time is the busy season. It is good also as carrying out the process which acquires and uses course ID preferentially. On the other hand, in the case where the user tends to avoid the busy season when using a course that has not been used in the past, when the current time is a non-busy season or a quiet season, the calculation unit 12 A process other than the technique based on the purchase history of the user may be preferentially used to perform the process of acquiring the course ID.

Referring to FIG. 7 again, in step S15, the calculation unit 12 acquires the plan ID associated with the course ID acquired in step S14 and the plan information as plan information candidates to be recommended to the user (S15). ). Specifically, the calculation unit 12 refers to the plan information storage unit 21, and among the plan IDs associated with the course ID acquired in step S14, for example, from the current date as a playable day. A plan ID and plan information associated with dates from one week to three weeks later are acquired.

Next, the calculation unit 12 calculates a score for each plan information candidate acquired in step S15 (S16). Specifically, the calculation unit 12 calculates a score based on at least one of the purchase history of the user and the plan information of the reference plan and the attribute of the plan information candidate.

For example, the calculation unit 12 calculates the score of the plan information candidate based on the similarity between the attribute of the plan information candidate and the purchase history or the attribute of the reference plan. The score calculated in this way has a higher value as the similarity is higher. This will be specifically described below.

For example, the calculation unit 12 calculates a score for each plan information candidate by the following equation (1).

The left side in Expression (1) represents the score of the plan information candidate. The first term on the right side represents the similarity of charges between the plan information candidate and the reference plan or user behavior data of the user. w _p is a predetermined coefficient and is set by design. The details of the first term on the right side can be expressed, for example, as in Expression (2).

S (price | price _ref , u) on the first side in Expression (2) represents the similarity between the plan information candidate charge and the charge of the reference plan or the user behavior data of the user. Further, t and t _ref mean that the similarity of time is taken into account in the calculation of the similarity.

As shown in the third side, the second term in the denominator of the second side of Equation (2) is the distance between the price of the plan information candidate and the price of the reference plan or the user behavior data of the user. To express. SD (price _user ) represents the standard deviation of the charge from the user behavior data. AVG (price) _t and AVG (price) _tref represent the average of the plan information candidate plan and the reference plan average, respectively, at the time of the plan to be calculated. The weekend _ratio is a coefficient for correcting the difference between the weekday charge and the weekend charge (Saturday, Sunday, public holiday). The plan information candidate plans are for weekends. If the person is a weekday target, a number greater than 1 (for example, 1.4) is set, the plan information candidate plan is for a weekday, and the reference plan is for a weekend. If it is, a number smaller than 1 (for example, the reciprocal of 1.4) is set. This coefficient is set empirically or empirically. Also, if both the plan information candidate plan and the reference plan are for weekdays, or if both the plan information candidate plan and the reference plan are for weekends, this coefficient is set to 1. Is done. “price” and “price _ref” are charges for the plan information candidate and the reference plan, respectively.

The second term on the right side of Expression (1) represents an optional similarity between the plan information candidate and the reference plan or user behavior data of the user. The option is the presence / absence of lunch, caddy, discount, etc. in the contents of the golf plan, and the opt is data of these presence / absence (for example, as vector data). w _k is a predetermined coefficient and is set by design. n is the number of optional items.

The third term on the right side of Expression (1) represents the similarity of the attribute regarding the course between the plan information candidate and the reference plan or the user behavior data of the user. The course is data (for example, as vector data) of attributes related to the course of the plan information candidate. The course _ref is obtained by converting the attribute regarding the course in the reference plan or the user behavior data of the user into data (for example, as vector data).

As described above, the plan information candidate is obtained by converting the plan information candidate and the attribute of the reference plan or the user behavior data of the user into data (for example, as vector data) and calculating the similarity (for example, the distance between the vectors). The score is calculated. As shown in Equation (1) and Equation (2), in this embodiment, the similarity for each element such as fee, various attributes of the plan, course, etc. is calculated, and the weight is adjusted by the coefficient. The score is calculated by adding the similarities. By calculating the score in this way, various elements are appropriately reflected in the score. Therefore, the plan recommendation based on the score is highly likely to meet the user's wishes. Note that the score calculation according to the expressions (1) and (2) is an example of the score calculation, and the similarity used for the score calculation can be calculated by various known techniques. By calculating the score in this way, the calculated score reflects the degree of plan information candidate user's desire.

In other words, the plan includes various elements such as fees, various attributes of the plan, and courses, but the conventional recommendation system recommends a golf course that is another course by focusing on the course that is one element. It was to stay in. This is because, as described above, collaborative filtering may not function effectively when the effective period such as a plan is limited. For this reason, conventionally, collaborative filtering by courses has been performed focusing on courses that exist for a long time and are considered to be important elements for users. However, the course is only one element of the plan. In practice, the user selects the plan by paying attention to the elements of the plan other than the course, or selects the plan by comprehensively judging a plurality of elements. It is thought that there is. Therefore, in this embodiment, as described above, the plan is decomposed into elements, the similarity between the reference plan and the plan information candidate is calculated for each element, and the overall score of the plan is calculated. . With such a score, it is possible to extract and recommend a plan that suits the user's preference while flexibly considering various factors.

Next, a second example of the score calculation process in step S3 will be described with reference to FIG. First, the calculation unit 12 refers to the user information storage unit 22 based on the user ID acquired in step S2, and acquires a user purchase history (S31). The calculation unit 12 also refers to the plan information storage unit 21 in order to acquire plan information of the plan included in the purchase history.

Next, in the plurality of plans included in the purchase history of the user, the calculation unit 12 determines whether or not there is a predetermined tendency with respect to the location associated with the plan, that is, the location where the golf course is located (S32). The predetermined tendency is, for example, a tendency that the user selects a course and a plan according to the distance to the golf course.

For example, in a plurality of plans included in the purchase history, the calculation unit 12 determines a predetermined location-related information when there is a correlation of a predetermined level or more between the distance from the user's location to the course and the start time of the play. It is determined that there is a tendency. The presence or absence of correlation can be determined by a well-known statistical process. In addition, for example, in the course associated with a plurality of plans included in the purchase history, the calculation unit 12 determines a predetermined place regarding the place when the ratio of the course at a certain distance from the user's location is equal to or greater than a predetermined degree. It may be determined that there is a tendency. Further, for example, the calculation unit 12 may determine that there is a predetermined tendency regarding the location when the standard deviation of the geographical parameter in the user behavior data of the user is equal to or less than a predetermined value.

Next, the calculation unit 12 calculates a score for each piece of plan information according to the presence or absence of a predetermined tendency regarding the place determined in step S32. (S33). The calculation of the score may be performed using Expression (1) and Expression (2) similarly to Step S16. In calculating the score, the calculation unit 12 sets weights for geographical parameters according to the presence or absence of a predetermined tendency regarding the place. The geographical parameter is, for example, a parameter related to the location among various parameters used for calculating the score, and is a parameter related to the location of the course, the location of the user, and the like. That is, when it is determined in step S32 that there is a predetermined tendency regarding the place, the calculation unit 12 uses a geographic parameter in calculating the score, compared with a case where it is not determined that there is a predetermined tendency. In some cases, the weighting of these parameters is increased. By calculating the score in this way, the user's desire for the plan is more appropriately reflected on the calculated score.

Next, a third example of the score calculation process in step S3 will be described with reference to FIG. First, the calculation unit 12 refers to the user information storage unit 22 based on the user ID acquired in step S2, and acquires the purchase history of the user (S41). The calculation unit 12 also refers to the plan information storage unit 21 in order to acquire plan information of the plan included in the purchase history.

Thus, for example, when a user selects a plan for a course that is located farther from the user's location, if there is a tendency to select a plan for a later time, such a tendency Even if the user purchases the plan only on Saturday and Sunday, if the plan of the night game exists in the course near the user's address or work location, the plan is presented. Control can be performed.

Subsequently, the calculation unit 12 acquires, as a reference plan (reference plan information), the plan ID of one plan information purchased in the past in the purchase history (S42). The process of step S42 is the same as the process of step S12. Next, the calculation unit 12 determines whether or not there is a predetermined tendency with respect to the location associated with the plan, that is, the location where the golf course is located, in the plurality of plans included in the purchase history of the user (S43). The process of step S43 is the same as the process of step S32.

Next, the calculation unit 12 acquires a course ID (reference location information) associated with the plan ID of the plan acquired as the reference plan (S44). The process of step S44 is the same as the process of step S13.

Subsequently, the calculation unit 12 acquires a course ID (location information candidate) of a recommended course based on the course ID of the reference plan (S45). Here, when it is determined in step S43 that there is a predetermined tendency related to the place, the calculation unit 12 performs course ID acquisition processing by weighting the geographical parameters. Specifically, for example, in the process of step S14 described with reference to FIG. 8, after the course ID is acquired based on the purchase history of the user (S21), the acquired course ID is less than a predetermined number. Although the course ID is acquired based on the geographical information, in step S45, the calculation unit 12 first acquires the course ID based on the geographical information, and then the number of acquired course IDs reaches a predetermined number. When there is no course ID, the course ID may be acquired by the user's purchase history or collaborative filtering.

Next, the calculation unit 12 acquires the plan ID associated with the course ID acquired in step S45 and the plan information as plan information candidates to be recommended to the user (S46). The process of step S46 is the same as the process of step S15.

Next, the calculation unit 12 calculates a score for each piece of plan information according to the presence or absence of the predetermined tendency related to the place determined in step S43. (S47). The process of step S47 is the same as the process of step S33.

As described with reference to FIGS. 9 and 10, by calculating the score in consideration of geographical conditions such as the course location, it becomes possible to make a recommendation that better meets the needs of the user. . Furthermore, the following examples can be given as recommendations that take geographical conditions into consideration.

For example, in the user's purchase history, in the correlation between the course location and the start time of the play, the longer the distance from the user's location to the course location, the more likely the plan is to select a plan with a slower start time. If there is, the calculation unit 12 increases the weighting of the geographical parameter in calculating the score.

Further, in the purchase history of the user, if there is a tendency to purchase a plan associated with a course whose distance from the user's location is within a certain distance, the calculation unit 12 is fixed from the user's location. In order to increase the score of the course within the distance, the weight of the course location is increased when calculating the score. In addition, the weight of the course location may be increased when calculating the score so that the score of the course in which the required time from the user location to the course location is within a certain time is high. On the other hand, for a course that is not within a certain distance from the user's location, the calculation unit 12 is similar to a plan with a late play start time or a course attribute other than the location in the purchase history of the user. Increase the score of such a plan so that the plan is recommended.

In addition, when there is a plan where the distance from the user's location or work place is within a certain distance and the play start time is slow (for example, after 18:00), such a plan is recommended to the user. As described above, the calculation unit 12 calculates the score by weighting the course location and the start time of the play so as to increase the score of such a plan.

In addition, in the purchase history of the user, when only a plan for a course located in a certain area (for example, within a prefecture) is purchased, the calculation unit 12 makes a recommendation so that the plan for the course in the area is recommended. , Increase the weight of the course location.

Refer to FIG. 6 again. In step S4, the transmission unit 13 of the recommendation device 10 transmits plan information including the score calculated in step S3 to the terminal 30 (S4). Note that the calculation unit 12 may extract plan information based on the calculated score, and cause the transmission unit 13 to transmit the extracted plan information. For example, the calculation unit 12 may extract a predetermined number of plans having a high calculated score from the top ones and cause the transmission unit 13 to transmit the plan information of the extracted plans. On the other hand, the plan information acquisition unit 32 of the terminal 30 acquires the plan information transmitted from the recommendation device 10 (S5). The plan information transmitted to the terminal 30 includes at least a play date, course (course ID) information (location information), and a score. All may be included.

Next, the display control unit 33 of the terminal 30 displays the plan information acquired in step S5 on the display unit of the terminal 30 (S6). For example, the display control unit 33 may extract a predetermined number of plan information from the acquired plan information in descending order of score, and display the extracted plan information on the display unit.

Further, the display control unit 33 may display the plan information acquired by the plan information acquisition unit 32 on a display column displayed on the display unit of the terminal 30 for each date in time series. An example of this display processing will be described with reference to FIG.

The display control unit 33 extracts plan information for each date from the plan information received in step S5 (S51). The display control unit 33 displays the plan information extracted for each date in the first display column for each date in time series. The first display field is configured in a calendar format, for example. The calendar represents dates in a tabular format. FIG. 12 is a diagram illustrating a display example of the plan information displayed in step S52. As shown in FIG. 12, the plan information is displayed for each calendar date. In addition, when there are a plurality of pieces of plan information associated with each date, the display control unit 33 associates the plan information with the highest score among the plurality of pieces of plan information associated with the same date with the date. The plan information selected at random from a plurality of pieces of plan information associated with the same date may be displayed in association with the date.

When the display control unit 33 displays the plan information in association with each date, the display control unit 33 displays plan information including attributes different from the attributes associated with the plan information displayed in association with the first date. Control may be performed so that the date is displayed in association with the date. This control will be described with reference to FIG. FIG. 13 is a diagram illustrating an example of plan information display control. Specifically, in FIG. 13, the display control unit 33 displays the plan information of the course “AAA country” in association with the date “January 14” (first date). At this time, the display control unit 33 displays the plan information of the course “BBB golf club”, which is an attribute relating to a place different from the course “AAA country”, in association with the date “January 15” (second date). Let Thus, since the plan information having an attribute different from the plan information displayed in association with the first date is displayed in association with the second date, the user can obtain various plan information. .

Next, the receiving unit 34 determines whether or not a plan selection input by the user has been received (S53). If a plan selection input is accepted, the process proceeds to step S54. In step S54, the display control unit 33 displays other plan information associated with the same date as the selected plan in a second display field different from the first display field (S54). Specifically, the display control unit 33 displays other plan information associated with the same date as the selected plan in another field (second display) different from the calendar display field (first display field). Column).

FIG. 14 is a diagram showing a display example of the plan information displayed in another column. As shown in FIG. 14, when the accepting unit 34 accepts a selection input for the plan information displayed in the column “January 15” in the calendar column C, the display control unit 33 displays the date “January 15”. Plan information associated with "day" is displayed in another column D.

Specifically, the display control unit 33 is plan information associated with the date “January 15”, and includes courses “AAA country”, “BBB golf club”, “EEE golf club”, “FFF country”. ”And“ GGG country ”are displayed in another column D in association with the five plan information respectively associated with the location attributes. Through such display control, plan information associated with the same date as the plan information selected by the user is presented to the user. As a result, the user can obtain a plurality of pieces of plan information associated with dates that can be used by the user.

Next, the accepting unit 34 determines whether or not a selection input from the user has been accepted for the plan information displayed in another field D which is the second display field (S55). If a plan selection input is accepted, the process proceeds to step S56. In step S56, the display control unit 33 displays the plan information selected by the user in another column D and the plan information in which the course that is the attribute of the same location is associated with the calendar column C as the first display column. Display by date.

FIG. 15 is a diagram showing a display example of the plan information displayed in the calendar column C. As illustrated in FIG. 15, when the receiving unit 34 receives a selection input for the plan information of the course “EEE Golf Club” in another field D, the display control unit 33 displays the course “EEE Golf Club” which is an attribute related to the place. ”And the plan information in which the dates“ January 13 ”to“ January 17 ”displayed in the calendar column C are associated as play date attributes are received from the recommendation device 10. Extracted from the transmitted plan information and displayed in each date column. By such display control, plan information having the same location attribute as the plan information selected by the user and associated with another date is presented to the user together with the selected plan information. Thereby, the user can obtain plan information regarding a place that meets his / her wishes for more dates.

Note that the display control unit 33 may perform display control of plan information based on the purchase history of the user. FIG. 16 is a diagram illustrating a display example of the plan information displayed based on the purchase history of the user. Specifically, the display control unit 33 determines whether or not there is a bias with respect to a specific day of the week for the plan information included in the purchase history of the user. The determination of the presence or absence of bias is realized based on a well-known statistical process. When it is determined that there is a bias with respect to a specific day of the week, the display control unit 33 causes the display unit to display a calendar-type display column that includes only the day of the week that has been determined to have a bias, and displays each display column. The plan information associated with the corresponding date is extracted from the plan information transmitted from the recommendation device 10 and displayed in each date column.

As shown in FIG. 16, the display control unit 33 causes the display unit to display a calendar-type display column composed only of Saturday and Sunday, and displays the dates “January 11” and “January 12” corresponding to the respective display columns. Plan information associated with “day”, “January 18”, “January 19”, “January 25”, and “January 26” is displayed in the respective columns. With such display control, it is possible to present plan information for a schedule that is likely to be purchased by the user. Further, by performing such display control, it is possible to reduce the date to be displayed, so even if the screen space for displaying the plan information to be recommended is limited, the space can be used by the user. Plan information that is highly likely to meet your wishes can be displayed.

As described above, the plan information display process in step S6 ends.

Next, a recommendation program for causing a computer to function as the recommendation system 1 will be described with reference to FIG. FIG. 17A is a diagram showing a recommendation program P10 for causing a computer to function as the recommendation device 10. The recommendation program P10 includes a main module m10, an acquisition module m11, a calculation module m12, and a transmission module m13.

The main module m10 is a part that comprehensively controls the recommendation process. The functions realized by executing the acquisition module m11, the calculation module m12, and the transmission module m13 are the same as the functions of the acquisition unit 11, the calculation unit 12, and the transmission unit 13 of the recommendation device 10 illustrated in FIG.

FIG. 17B is a diagram showing a terminal recommendation program P30 for causing a computer to function as the terminal 30. The terminal recommendation program P30 includes a main module m30, a user ID transmission module m31, a plan information acquisition module m32, a display control module m33, and a reception module m34.

The main module m30 is a part that comprehensively controls the recommendation processing in the terminal 30. The functions realized by executing the user ID transmission module m31, the plan information acquisition module m32, the display control module m33, and the reception module m34 are respectively the user ID transmission unit 31 and the plan information acquisition unit of the terminal 30 shown in FIG. 32, the same functions as those of the display control unit 33 and the reception unit 34.

The recommendation program P10 and the terminal recommendation program P30 are provided by storage media D10 and D30 such as a CD-ROM, a DVD-ROM, or a semiconductor memory, for example. Further, the recommendation program P10 and the terminal recommendation program P30 may be provided via a communication network as computer data signals superimposed on a carrier wave.

According to the recommendation system 1, the recommendation method, and the recommendation program of this embodiment described above, the plan extracted based on the user's history is displayed for each date, so the plan is presented to the user for each date. Is done. Thereby, the user can browse a plan associated with an available date. Therefore, the user can obtain useful plan information. In addition, a score for each plan is calculated based on the history of purchase or viewing of the user's plan, and plan information is presented to the user based on the calculated score. The score calculated based on the user's history is highly likely to reflect the degree of user's desire. Since the plan information is presented to the user based on such a score, the user can obtain the plan information of the plan that meets his / her wish.

The present invention has been described in detail above based on the embodiments. However, the present invention is not limited to the above embodiment. The present invention can be variously modified without departing from the gist thereof.

In this embodiment, the display control unit 33 is provided in the terminal 30, but the display control unit 33 may be provided in the recommendation device 10. In this case, the plan information is displayed on the display unit of the terminal 30 in various manners based on the display control of the display control unit in the recommendation device 10.

In this embodiment, in step S14 of the flowchart of FIG. 7 and step S45 of the flowchart of FIG. 10, as a method for acquiring the course ID, a method based on the purchase history of the user, a method based on collaborative filtering, a golf course, A method based on the geographical information of the user is exemplified. As a method for acquiring the course ID, the following method may be used.

For example, the calculation unit 12 refers to the purchase history of the user as a variation of the method based on the golf course and the geographical information of the user, acquires the position of the golf course associated with the purchased plan, An average position of the acquired positions of the plurality of courses is calculated, and a course ID of a course located at a position close to the calculated average position is acquired. Here, the calculation unit 12 may further acquire the location of the user, and obtain a course ID by weighting the course closer to the user's location as seen from the calculated average position. Further, the calculation unit 12 may perform correction to move the calculated average position by a predetermined distance in the direction of the user's location, and acquire the course ID based on the corrected average position.

Further, the calculation unit 12 may use a technique of calculating a score for each course by applying a so-called page rank (registered trademark) algorithm and acquiring a course ID of a course having a high score. For example, a user who has a page rank as a user, a link in the page rank as a participation in a plan purchased by another user, and a user who has purchased a plan in which more users have participated has a high importance and a user with a high importance. It can be considered that the course of the plan selected by is highly important. Specifically, if a user X makes a reservation for a plan associated with course A and three other users (each having a score of 1) participate in the plan, Thus, 3 points, which are the total points of other participating users, are given. Similarly, if the user Y reserves a plan associated with the course B and other users including the user X participate in the plan, the three points that the user X has for the user Y and others The total points of the participating users are given. Furthermore, if user Z makes a reservation for a plan associated with course A and another user participates in the plan, similarly, the total of other users' points will be given to user Z. It is done. In such a case, the calculation unit 12 calculates, as the score of the course A, the total value of the points of the user X, the user Z, and so on who have reserved the plan associated with the course A. The calculation unit 12 may extract and acquire course IDs of a predetermined number of higher-order courses based on the order of scores calculated for each course in this way. Moreover, the calculation part 12 is good also as acquiring the course ID of the course extracted based on a course and the user's geographical information further with respect to the course acquired in this way.

Further, the calculation unit 12 may acquire the course ID by so-called content-based filtering. Specifically, the calculation unit 12 acquires the course ID of the reference plan, for example, by the process shown in step S13 of the flowchart of FIG. The calculation unit 12 refers to the course information that stores the characteristics of each course in association with each other, and has attribute information similar to various attribute information associated with the acquired course ID. Extract the course. The similarity of the course features can be calculated by a known technique. For example, the attribute information for each course is expressed as a vector, and can be calculated as a distance between the vectors. And the calculation part 12 acquires course ID of the acquired other course.

Also, when the recommendation system 1 is linked to a social networking service (SNS) system and can acquire various information from the SNS system, the course ID can also be acquired using the user's network. For example, specifically, the calculation unit 12 acquires information on other users related to the user from the SNS system, refers to the purchase history of the other users, and purchases or browses the plans purchased by other users. The course ID of the associated course is acquired as the course ID used for recommending the plan to the user.

DESCRIPTION OF SYMBOLS 1 ... Recommendation system, 10 ... Recommendation apparatus, 11 ... Acquisition part, 12 ... Calculation part, 13 ... Transmission part, 21 ... Plan information storage part, 22 ... User information storage part, 30 ... Terminal, 31 ... User ID transmission part, 32 ... Plan information acquisition unit, 33 ... Display control unit, 34 ... Reception unit, D10, D30 ... Storage medium, m10 ... Main module, m11 ... Acquisition module, m12 ... Calculation module, m13 ... Transmission module, m30 ... Main module, m31 ... user ID transmission module, m32 ... plan information acquisition module, m33 ... display control module, m34 ... reception module, N ... network, P10 ... recommendation program, P30 ... terminal recommendation program.

Claims

A recommendation system that recommends plans to users,
Obtaining means for obtaining a user ID for identifying a user;
Based on the history of plan selection by the user specified by the user ID, at least one method among a plurality of methods for extracting a plan to recommend to the user is determined, and the plan is extracted using the determined method Extraction means to
Display control means for displaying the plan information of the extracted plan;
A recommendation system with
The extraction means includes
Based on the history of the user, including a calculation means for calculating a score for each plan,
Extracting a plan based on the score calculated by the calculating means;
The recommendation system according to claim 1.
The calculating means includes
In the history, one plan information selected in the past is acquired as reference plan information,
Based on the reference location information for specifying the location associated with the plan of the reference plan information, referring to the plan information storage means storing the plan information, a candidate to recommend to the user by a predetermined method To get candidate location information
The plan information associated with the acquired location information candidate is acquired as a plan information candidate to be recommended to the user,
For each acquired plan information candidate, a score is calculated based on at least one of the history and the reference plan information and the attribute of the plan information candidate.
The recommendation system according to claim 2.
The calculation means obtains the plan information of the latest selected plan in the history as the reference plan information.
The recommendation system according to claim 3.
The calculating means includes
Of the locations associated with the plan included in the history, location information of a location satisfying a predetermined condition is acquired as the location information candidate, and the number of the location information candidates acquired based on the history is preset. The location information candidate is acquired based on at least one of collaborative filtering based on the user's history and geographical information associated with the user when the predetermined number is not satisfied.
The recommendation system according to claim 3 or 4.
The calculating means includes
In the history of the user, when the number of selected plans is a predetermined number or more, the location information of the location satisfying a predetermined condition among the locations associated with the plan included in the history is the location. As information candidates,
In the history of the user, when the number of selected plans is less than a predetermined number, at least one of collaborative filtering based on the user history and geographical information associated with the user Based on the candidate location information,
The recommendation system according to claim 3 or 4.
The calculating means includes
The location information candidate is obtained by each of a plurality of predetermined methods,
For each of the predetermined methods, the acquisition of the plan information candidate associated with the location information candidate, and the calculation of the score of each plan information candidate,
The extraction means includes
For each of the predetermined methods, the plan is extracted from the plan information candidates based on the calculated score,
A predetermined method having the highest degree of matching with the plan included in the history of the extracted plan is determined as a method for extracting a plan to recommend to the user.
The recommendation system according to claim 3 or 4.
The calculation means calculates a score of the plan information candidate based on a similarity between the attribute of the plan information candidate and the attribute of the history or the reference plan information, and the calculated score has a high similarity So high,
The recommendation system according to any one of claims 3 to 7.
The calculating means includes
The presence / absence of a predetermined tendency is determined for the location associated with the plan included in the history of the user, and if it is determined that there is the predetermined tendency, it is not determined that there is the predetermined tendency If a geographic parameter is used in calculating the score for each plan, the weight for the parameter is increased.
The recommendation system according to any one of claims 2 to 8.
The plan information is plan information relating to golf play, and includes information on a golf course as an attribute relating to a place, and information relating to a date and time of playing as an attribute relating to a date.
The recommendation system according to any one of claims 1 to 9.
A recommendation method in a recommendation system for recommending a plan to a user,
An acquisition step of acquiring a user ID identifying the user;
Based on the history of plan selection by the user specified by the user ID, at least one method among a plurality of methods for extracting a plan to recommend to the user is determined, and the plan is extracted using the determined method An extraction step to
A display control step for displaying the plan information of the extracted plan;
A recommendation method comprising:
A recommendation program for causing a computer to function as a recommendation system for recommending a plan to a user,
In the computer,
An acquisition function for acquiring a user ID for identifying a user;
Based on the history of plan selection by the user specified by the user ID, at least one method among a plurality of methods for extracting a plan to recommend to the user is determined, and the plan is extracted using the determined method An extraction function to
A display control function to display the plan information of the extracted plan;
A recommendation program that realizes